A Rule-based Spelling Checker for Correcting Punctuation Errors in Indonesia Text using KEBI 1.0 Checker

Abstract

Punctuation errors in Indonesian often occur in Indonesian scientific writing. This error was caused intentionally and unintentionally by the author. The factor of intentional is because the author does not have time to make independent improvements after writing, while unintentional occurs because of the author's ignorance of the appropriate form of punctuation. Therefore, the punctuation error detection system in Indonesian writing was developed to overcome them. The name of the application is KEBI 1.0 Checker which was a web-based application and provided three features, namely punctuation detection, typo detection, and standard words detection. The research proposed to evaluate the performance of application based on punctuation errors using a rule-based spelling checker that can be detected were in the form of comma punctuation, hyphens in repeat words, and writing of pre-, post-, and inter- particles, as well as writing prefixes with additional English verbs. After the application was tested on ten scientific papers, KEBI 1.0 Checker showed accuracy for correcting punctuation errors reached an accuracy of 76.67%, precision of 100%, and recall of 76.67%.